calibre2jellyfin

Python script to construct a Jellyfin ebook library from a Calibre library.

Linux

(The Windows version of this script is maintained as a separate project. Please see wincalibre2jellyfin.)

Overview

• Created folder structure (foldermode in .cfg) is one of:
  • .../author/series/book/...
  • .../series/book/...
  • .../book/...
• Destination book folder contains
  • symlink to (single) book file in Calibre library (based on configured order of preference)
  • symlink to cover image in Calibre library
  • copy, possibly modified, of Calibre's metadata file
• Books may be selected in the .cfg file by author folder or by subject.
  • Allows you to exclude from Jellyfin those messy persistent remnants from the years when your library was scattered over multiple proprietary platforms.
  • Allows you to separate differing levels of mature content into separate Jellyfin libraries, access to which can be restricted within Jellyfin. (See Real Life section below for further discussion.)
  • Allows you to dispose books into libraries with differing structures as called out below in Series Handling
• Alternatively all books in the source library may be exported.
• Series handling
  • Foldermode is author/series/book
    • Suitable for fiction libraries
    • The script will attempt to extract series and series index from Calibre's metadata file.
    • If found, the target book folder name will be prepended with the series index. Optionally, the metadata <dc:title> and the <meta name="calibre:title_sort" content="...sort title..."> may be treated in the same way.
    • A short header identifying the index and series is prepended to the book description.
    • If series info is expected but not found, the structure collapses to .../author/book/.... and no mangling is performed.
  • Foldermode is series/book
    • Suitable for eComic libraries
    • This mode is similar to author/series/book above except there is no grouping by author, only by series and book, unless the series info is missing in which case the structure collapses to .../book/...
  • Foldermode is book
    • Suitable for non-fiction libraries
    • Books are organized strictly by book title
• Multiple output libraries may be configured

Example author/series/book structure

Example assumes script has been configured to prefer .epub types over .azw and .mobi.

└── Author/
    └── Book A/
    │   ├── cover.jpg
    │   ├── metadata.opf
    │   ├── Book A.azw
    │   └── Book A.epub
    ├── Book B/
    │   ├── cover.jpg
    │   ├── metadata.opf
    │   ├── Book B.mobi
    │   └── Book B.epub
   

└── Author/
    └── Lorem ipsum dolor sit amet Series/
        ├── 001 - Book A/
        │   ├── cover.jpg      <- symlink
        │   ├── metadata.opf   <- modified copy
        │   └── Book A.epub    <- symlink
        ├── 002 - Book B/
        │   ├── cover.jpg      <- symlink
        │   ├── metadata.opf   <- modified copy
        │   └── Book B.epub    <- symlink
   

Example series/book structure

The "series/book" option is intended for use with eComics, thanks for this go to Cudail.

└── Author M/
    └── Comic A/
        ├── cover.jpg
        ├── metadata.opf
        └── Comic A.cbz
└── Author N/
    └── Comic B/
        ├── cover.jpg
        ├── metadata.opf
        └── Comic B.cbz
   

└── Lorem ipsum dolor sit amet Series/
    ├── 001 - Comic A/
    │   ├── cover.jpg      <- symlink
    │   ├── metadata.opf   <- modified copy
    │   └── Comic A.cbz    <- symlink
    ├── 002 - Comic B/
    │   ├── cover.jpg      <- symlink
    │   ├── metadata.opf   <- modified copy
    │   └── Comic B.cbz    <- symlink
   

Changes

• 2026-03-17 (Current version, Branch Main)
  • New configuration parameter
    • additionalAuthors = [0|1]
    • Causes books to appear in folders for additional authors as well as the "primary" author.
    • Applies only when foldermode is 'author,series,book'
    • Add the new parameter to your configuration file as desired.
• 2024-12-14
  • Added option to configuration parameter mangleMetaTitle to allow prefixing metadata title with unformatted series index (e.g. "7" instead of "007"). Run once with --update-all-metadata to apply.
  • Added --invert command line option
    • This causes the output of --list to show excluded items. Books having a type that is not in a configured bookfiletypes parameter are ignored. I.e. such books are considered to be neither selected nor excluded.
  • --list reports now separate output by source library
  • Added additional INFO logging
• 2024-11-22
  • Added support for selection of all books in the source library
  • Added support for Selection of books by subject (aka tags in Calibre)
  • Added command line options
    • --dryrun
      • This displays normal console output plus showing where files would be output but does not actually export anything.
    • --debug
      • Displays lots of path and metadata information. Useful if you need to look into why a book is or is not being selected.
    • --version
    • --list LIST_SPEC
      • Outputs a report that can be useful in curating your library.
  • New configurations (see example .cfg)
    • selectionMode = [author | subject | all]
    • subjects = ...
  • Additional warnings
    • Missing cover file
    • Missing metadata file
      • Note: In 'subject' selection mode, if the metadata is missing the book cannot be exported.
  • All authors now appear in the description header and the authors by-line appears in all folder modes whether there is series info or not.
    • Run once with --update-all-metadata to apply
  • Upgrade considerations
    • If you already have the 2024-09-02 version and are not interested in selection by subject, there is not much reason to download this version. The selection by author functionality has not changed.
    • If you download this version because you are interested in the new warnings or the new command line options but not selection by subject or selection of all books, you do not need to make any changes to your configuration file. Behavior defaults to selection by author.
    • If you are interested in selection by subject, you will need to add the selectionMode and subjects parameters to your configuration file. In 'author' or 'all' selectionMode, any subjects list is ignored. In 'subject' or 'all' selectionMode, any authors list is ignored. You can maintain both lists and switch between them using selectionMode.
    • If you are interested in selection of all books, you will need to add just the selectionMode parameter.
• 2024-09-02
  • Add author's name to book description.
  • Add support for fractional series indices maintaining sort
    • '' -> '999'
    • '3' -> '003'
    • '34' -> '034'
    • '345' -> '345'
    • '3456' -> '3456'
    • '3.2' -> '003.02' <- new
    • Notes
      • Any series books that had fractional indices prior to this version will appear as new book folders with the index formatted in the new way and leaving the prior formatted versions as duplicates. You will probably want to delete, in Jellyfin, the old versions.
  • Add warning when book file of configured type not found in book folder
• 2024-06-19
  • Add support for "series/book" mode, thanks to Cudail
• 2024-02-21
  • Coding/style improvements and lint cleanup only. No functional changes. If you already have the 2024-01-27 version installed there is no real reason to download this version.
• 2024-01-27
  • Add support for mangling the metadata title sort value
  • Make metadata mangling behavior configurable (new configuration parameters)
    • mangleMetaTitle = [1 | 0]
    • mangleMetaTitleSort = [1 | 0]
  • Add command line option to force update of all metadata files.
    • --update-all-metadata

Dependencies

• Python 3

Installation

1. In your browser navigate to "https://github.com/shawn61cp/calibre2jellyfin"
2. Click the green "Code" button
3. In the resulting dropdown, just over halfway down, find and click on "Download ZIP".
4. Save the zip file somewhere convenient and extract it.  We will call this EXTRACT_FOLDER.
5. Change to directory EXTRACT_FOLDER/calibre2jellyfin-main
6. In a terminal:
7.      $ chmod 755 calibre2jellyfin.py
8. Choose a location to install the script.  You may want to add this location to your path.  We will call this INSTALL_FOLDER.
9. In a terminal:
10.     $ cp calibre2jellyfin.py INSTALL_FOLDER/
11.     $ cp calibre2jellyfin.cfg ~/.config/

Usage

• Edit the ~/.config/calibre2jellyfin.cfg file before first use.
  • Following the comments in the .cfg, set up your source and destination storage locations (libraries), your folder mode, your list of authors, and title mangling.
• If you added the calibre2jellyfin.py script to a location on your path, in a terminal:
  • $ calibre2jellyfin.py
• If not, include the full path to the script:
  • $ INSTALL_FOLDER/calibre2jellyfin.py

Upgrading

Two things need to be accomplished:

1. Replace your current script, wherever it was originally installed, with the new one.
   • This can be done basically by following installation steps 1 - 10. Do not perform step 11 since that would destroy your current configuration.
2. Add any new config options to your existing configuration file.
   • This can be done by copying and pasting any new configuration parameters from the new sample configuration into your current configuration, or even just editing your current configuration. New configuration options are listed in the Changes section and also in the sample .cfg file.

Command line options

usage: calibre2jellyfin.py [-h] [--debug] [--dryrun] [--invert] [--list LIST_SPEC]
                           [--update-all-metadata] [-v]

A utility to construct a Jellyfin ebook library from a Calibre library. Configuration file
"/home/shawn/.config/calibre2jellyfin.cfg" is required.

options:

  -h, --help            show this help message and exit
  
  --debug               Emit debug information.
  
  --dryrun              Displays normal console output but makes no changes to exported libraries.
  
  --invert              Inverts the sense of the --list argument, showing those items that will not be
                        exported. Only valid in combination with --list.
                        
  --list LIST_SPEC      Suspends normal export behavior. Instead prints info from configuration
                        sections and file system that is useful for curation. LIST_SPEC is a comma-
                        delimited list of columns to include in the report. The output is tab-
                        separated. Columns may be one or more of authors, section, book, bfolder,
                        afolder, subject, series, or index. authors: display author name if the source
                        folder exists. section: display section name. book: display book title.
                        bfolder: display book folder. afolder: display author folder. subject: display
                        subject that matched. series: display name of the series. index: display series
                        index. The report output is sorted so there will be a pause while all
                        configured sections are processed.
                        
  --update-all-metadata
                        Useful to force a one-time update of all metadata files, for instance when
                        configurable metadata mangling options have changed. (Normally metadata files
                        are only updated when missing or out-of-date.)
                        
  -v, --version         Display version string.
  

Real Life

The installation and usage instructions above work fine but other situations may be encountered or other conveniences desired.

Permissions on created Jellyfin library

The usage/installation steps described above yield a Jellyfin store that is owned by whatever user ran the script. Jellyfin can serve up this library because most default user configurations under Linux create files as world-readable. However, if you make the library owned by Jellyfin (the jellyfin Linux user/service account), you will be able to delete books through the Jellyfin interface assuming you have administrative permission on the library within the Jellyfin app itself.

If your situation is like mine, cleaning up my Calibre library, ensuring the right metadata and covers got downloaded etc., is an ongoing task. If you discover an issue while browsing your books within Jellyfin, you can delete the Jellyfin book (within Jellyfin), go back to your Calibre library, clean things up, and then re-run the script. Et voilà! Depending on the issue, you might not even have to delete the book from Jellyfin; Rerunning the script might be sufficient. However, deleting the book, or even the author folder if there were many changes, from Jellyfin does guarantee a clean re-creation.

Note 2025-02-24: I have found that Jellyfin will delete an entire containing author or other folder and all contents, but attemtping to delete at the book level results in only the book file itself
being deleted, along with I believe its internal metadata, but not the enclosing book folder
nor the cover image nor the metadata opf file. The book folder must be deleted through the file system. Since this is a bit of a nuisance, if I have very much to clean up, I sometimes just delete the author folder and let the whole author structure be re-created on the next run or via a manual run.

Note that because the book and cover files are soft linked, and the folders and metadata file are copies, when you delete a book or author through the Jellyfin interface, you are only affecting the Jellyfin library and not your precious Calibre library.

For myself, I arrange to run the script under the jellyfin account. This will result in the files and folders output by the script being owned by jellyfin. The default home directory for the jellyfin user is /var/lib/jellyfin. I create the path and install the script to /var/lib/jellyfin/.local/bin/calibre2jellyfin.py. Similarly I install the configuration file to /var/lib/jellyfin/.config/calibre2jellyfin.cfg.

Set up the script under the jellyfin account

The steps to accomplish the above follow. I refer again to the EXTRACT_FOLDER as described in the installation section above. Finally, jellyfin by default does not permit logins so these steps will be performed as root.

Change to root: $ sudo su -

Create the paths for the script and its configuration # mkdir -p /var/lib/jellyfin/.local/bin # mkdir -p /var/lib/jellyfin/.config

Copy the script and configuration. Note: If you set up a .cfg earlier, copy that instead of the sample .cfg from the EXTRACT_FOLDER. # cp EXTRACT_FOLDER/calibre2jellyfin-main/calibre2jellyfin.py /var/lib/jellyfin/.local/bin/ # cp EXTRACT_FOLDER/calibre2jellyfin-main/calibre2jellyfin.cfg /var/lib/jellyfin/.config/

Change ownership of the files and paths created above to jellyfin. # chown -R jellyfin:jellyfin /var/lib/jellyfin/.local # chown -R jellyfin:jellyfin /var/lib/jellyfin/.config

If you did not do so during the installation steps, make the script executable.
# chmod 755 /var/lib/jellyfin/.local/bin/calibre2jellyfin.py

Now, if you did not already have a .cfg set up, continue as root and edit the configuration file /var/lib/jellyfin/.config/calibre2jellyfin.cfg as described under Usage above.

Finally, exit from the root shell. # exit

Running the installed script

I find that when dealing with accounts that do not permit login that the runuser utility is convenient. Possibly this is habit. :) To run the script I execute the following from my own account. $ sudo runuser -u jellyfin -- /var/lib/jellyfin/.local/bin/calibre2jellyfin.py

Scheduling the script to run automatically

You can use CRON to run the script regularly to keep your Jellyfin library updated. The steps below assume the script has been set up under the jellyfin user as described above. The cron job is set up under the root account but again uses runuser to execute it as jellyfin.

Change to the root account $ sudo su -

Start the cron editor # crontab -e

Add a line like this to the cron file. The redirection causes any error messages to be included in the mailed log (if you have that enabled). The example below runs the script every night at 11:00 PM. 0 23 * * * runuser -u jellyfin -- /var/lib/jellyfin/.local/bin/calibre2jellyfin.py 2>&1

Save the file, exit the editor, and exit the root shell.

Calibre Author Folders

If you find that an expected author does not show up in the created Jellyfin library, double check that the author as listed in the .cfg file matches the actual folder name in the Calibre library.

In the case of multiple authors, Calibre expects them to be delimited by an ampersand and, I believe, a space on either side (" & "). If you acquired metadata with multiple authors delimited by commas, replace the commas with space-ampersand-space, but keep commas that separate name suffixes like Jr./Sr./PhD/MD etc. Using ampersands as delimiters accomplishes two things. First, Calibre will recognize each author and list them properly and individually in its author browser. Second, Calibre will name the author folder after the first author in the list, keeping your on-disk Calibre and Jellyfin libraries nice and clean, as opposed to the lengthy run-on name that results otherwise.

Another thing I have encountered is when multiple versions of the author name exist, such as "Public, John Q." and "John Q. Public", and they are then consolidated, Calibre actually moves the books into the folder matching the consolidated name. If the author name configured for calibre2jellyfin happened to match the author name that "went away", updates to that author's books may appear to die, or you might see two different versions of the same author in Jellyfin. The solution is to just delete, in Jellyfin, one or both authors, ensure that the author configured in the .cfg file matches the Calibre author folder, then re-run the script to have them cleanly re-created. Jellyfin will eventually detect the changes and update the display contents. You can also right-click on an item within Jellyfin and request an immediate metadata refresh. Even so sometimes it will take a few minutes for Jellyfin to recognize the changes.

Tricks with [Construct] jellyfinStore

Although the instructions in the example .cfg file state categorically that the jellyfinStore parameter should be set to the location of the Jellyfin library, there actually is some wiggle room.

Suppose that you want top level folders in your Jellyfin library that separate your fiction library into "Science Fiction", "Fantasy", "Westerns", and "Romance". You could create the following structure for the Jellyfin library and then create separate [Construct] sections for each top level category.

.../fiction/                <- point the Jellyfin library here
    ├── Fantasy/            <- point a [ConstructFantasy] jellyfinStore param here
    ├── Romance/            <- point a [ConstructRomance] jellyfinStore param here
    ├── Science Fiction/    <- point a [ConstructSciFi] jellyfinStore param here
    └── Westerns/           <- point a [ConstructWesterns] jellyfinStore param here

Then using your desired selectionMode, arrange for appropriate books to be output from each [Construct...] section. Jellyfin would then display drillable category folders above the author folders (or whatever folderMode you choose).

Mature Content

Selection by author, although good for this purpose, is not exactly 100% perfect since it is possible for a single author to write content of differing level.

Selection by subject gives finer grained control but tags are often missing and when present seem inconsistent. To really make selection-by-subject work you would probably have to curate tags yourself.

Another approach would be to combine construction methods. Nothing prevents having multiple [Construct] sections that output to the same Jellyfin library. (You probably would want to use the same foldermode.) One could exclude problematic authors (No offense, authors!) from a selection-by-author [Construct] and then handle those within a selection-by-subject [Construct].

Curation

None of these reports and lists are at all required in order to use the calibre2jellyfin script. I use them when I review my Calibre library for things that need to be cleaned up.

Caveat Usor: Several of following procedures use sqlite3 to access the Calibre metadata database directly. Read-only select statements should not present problems. Nevertheless it is a good idea to make a backup of such an important file.

Listing Calibre author folders etc. that will not be output by calibre2jellyfin.

The --list option together with the --invert option report items that will not be exported. I use this as my to-do list.

Compact list of Calibre author's series

sqlite3 -separator $'\t' PATH_TO_CALIBRE_LIBRARY/metadata.db '
select 
          A.name as author
        , S.name as series
from
                                      authors A
        inner JOIN                    books_authors_link BAL       on  BAL.author = A.id
        inner JOIN                    books_series_link BSL        on  BSL.book = BAL.book
        inner JOIN                    series S                     on  S.id = BSL.series
group by
          A.name
        , S.name
order by
        1, 2
;' | column -t -s $'\t' | less

Compact list of Calibre series

# With authors
sqlite3 -separator $'\t' PATH_TO_CALIBRE_LIBRARY/metadata.db "
select distinct
      S.name as series
    , (
        select group_concat(A.name, ',')
        from
                        books_authors_link BAL
            inner join  authors A                   on  A.id = BAL.author
        where 
            BAL.book = B.id
    ) as authors
from
                        books B
    inner join          books_series_link BSL       on  BSL.book = B.id
    inner join          series S                    on  S.id = BSL.series
order by
    1, 2
;" | column -t -s $'\t' | less

# With author folder
sqlite3 -separator $'\t' PATH_TO_CALIBRE_LIBRARY/metadata.db "
select distinct
      S.name as series
    , substr(B.path, 1, instr(B.path, '/')-1) as afolder
from
                        books B
    inner join          books_series_link BSL       on  BSL.book = B.id
    inner join          series S                    on  S.id = BSL.series
order by
    1, 2
;" | column -t -s $'\t' | less

Compact list of Calibre author's books

sqlite3 -separator $'\t' PATH_TO_CALIBRE_LIBRARY/metadata.db "
select
      A.name as author
    , B.title as book
    , coalesce(S.name, '') as series
from
                        authors A
    inner join          books_authors_link BAL      on  BAL.author = A.id
    inner join          books B                     on  B.id = BAL.book
    left join           books_series_link BSL       on  BSL.book = B.id
    left join           series S                    on  S.id = BSL.series	
order by
    1, 2
;" | column -t -s $'\t' | less

Compact list of Calibre collaborator's books

sqlite3 -separator $'\t' PATH_TO_CALIBRE_LIBRARY/metadata.db "
select
    (
        select group_concat(A.name, ',')
        from
                        books_authors_link BAL
            inner join  authors A                   on  A.id = BAL.author
        where 
            BAL.book = B.id
    ) as authors
    , B.title as book
    , coalesce(S.name, '') as series
from
                        books B
    left join           books_series_link BSL       on  BSL.book = B.id
    left join           series S                    on  S.id = BSL.series
order by
    1, 2
;" | column -t -s $'\t' | less

Differences between Linux and Windows versions

The Windows and Linux code is probably 99 point something percent identical but there are a few differences.

• Configuration file
  • Linux
    • Named: calibre2jellyfin.cfg
    • Resides in the .config/ directory under whatever account runs the script
  • Windows
    • Named: wincalibre2jellyfin.cfg
    • Resides in the same location as the script
• UTF-8 decode error
  • Linux
    • None
  • Windows
    • In testing the Windows version, 2 of 4000 books exhibited a utf-8 decode error of the metadata. Testing was performed on the same library under Windows as under Linux, although for Windows access was via a Samba share, while under Linux the library resided on a local ext4 file system.
    • The Windows version implements a fallback to backslashed ascii representation and logs a warning if utf-8 decoding fails.
• Symlinks
  • Linux
    • Creates symlinks for the book and cover image. The metadata file is copied since it may be modified.
  • Windows
    • Creates copies of all files
    • Windows supports symlinks but requires elevated privileges to create them. It seems inappropriate to me to run a script of this nature with administrative privileges.
    • In, admittedly shallow, testing Jellyfin does not read Windows shortcuts, not that I would expect it to since shortcuts are I believe an artefact of the shell, but I did look at the possibility. :)
    • I welcome advice from anyone more knowledgeable about these details than I.

Odds and Ends

• I have noticed that Jellyfin does not re-paginate if you resize the browser window or change the zoom factor after you have opened the book. However if you do these before opening the book it does so nicely.